From mike.duigou at oracle.com Sat Mar 15 00:19:55 2014 From: mike.duigou at oracle.com (Mike Duigou) Date: Fri, 14 Mar 2014 17:19:55 -0700 Subject: A webrev patch to use 'trees' instead of 'forest' In-Reply-To: <5321C040.1070204@oracle.com> References: <20140310102153.35176@eggemoggin.niobe.net> <532127D9.3050203@oracle.com> <5321C040.1070204@oracle.com> Message-ID: <7F1B9FD8-35D1-44CF-9CB4-FFC4923AEDC6@oracle.com> Hi Daniel; Since it looks like several people have tried your patch and seem to find it works for them then perhaps you should just create a CODETOOLS issue and propose it to this list (webrev-dev) for sponsorship. I haven't had a chance to try it yet but likely wouldn't have anything to add by my review (and I am not a committer/reviewer on CODETOOLS) so I would encourage you to go ahead with proposing it yourself. Cheers, Mike On Mar 13 2014, at 07:27 , Daniel Fuchs wrote: > Hi, > > [removed the jdk_confidential wider diffusion] > > If that can help, attached is a patch that modifies webrev.ksh > to use trees instead of forest. > > The option remains the same - you still call webrev -f but it uses > trees underneath. > > Note that for this to work you need to have your forest configured > for trees. An easy way to do that is to call: > > hg tconfig --set --walk --depth > > at the root of the forest. > > Disclaimer: > 1. I'm not an expert > 2. I tested the patch only on my Mac, and only with the > following options: > > [all files edited, no commit]: > > webrev -fN > > [all files committed, no edits]: > > webrev -f > > The main changes (trickiest) comes from the fact that whereas > hg f used to print '[]' hg t now > prints '[]:' > > Hope this helps, > > -- daniel > > > > On 3/13/14 7:16 AM, Mike Duigou wrote: >> You will have to remove the "-f" option. If you need to generate cross-repo webrevs then you'll need to create a filelist of the files and pass that to webrev. I typically use something like: >> >> cat filelist | grep -v generated | ksh ~/code/jdk/webrev/webrev.ksh -m -u mduigou -c `hg qtop` -r qparent -o ~/`hg qtop`/3 - >> >> (where the final trailing "-" is the "read stdin" alternative to providing a filename). >> >> I will make sure the "-f" option disappears in the next webrev revision. >> >> Mike >> >> >> On Mar 12 2014, at 20:36 , Lana Steuck wrote: >> >>> Was webrev update to work with the new hg (without forest commands)? I just cloned the latest webrev from http://hg.openjdk.java.net/code-tools/webrev , however it seems that it still uses the old forest command (foutgoing, fstatus, etc.), so it does not work with newer hg (unless I am missing something here...). >>> >>> >>> ssteuck at jcia-sparcv9:jdk7u-clone$ /net/jano1.us.oracle.com/export1/jcg/ws-depot/tools/webrev/webrev.ksh -m -f -o ../logs/ >>> SCM detected: mercurial >>> WARNING: parent rev is not tip. Maybe an update or merge is needed >>> File list from hg foutgoing ... >>> hg: unknown command 'foutgoing' >>> >>> No outgoing, perhaps you haven't commited. >>> File list from hg fstatus -mard ... Done. >>> Workspace: /export/home/jcia-jobs/7u-jdk7u-dev-pit.20140312-141606/jdk7u-clone >>> Output to: /export/home/jcia-jobs/7u-jdk7u-dev-pit.20140312-141606/jdk7u-clone/../logs//webrev >>> Output Files: >>> abort: cannot follow file not in parent revision: "Mercurial" >>> Mercurial >>> *** Error: file not in parent or child >>> abort: cannot follow file not in parent revision: "basic" >>> basic >>> *** Error: file not in parent or child >>> >>> >>> On 03/10/2014 10:21 AM, mark.reinhold at oracle.com wrote: >>>> All servers and up and functioning normally with the exception of >>>> mail.openjdk.java.net, which is working but slowly. >>>> >>>> Reminders about the Mercurial changes: >>>> >>>> - The push-only "-gate" repos are no longer be needed and have been >>>> removed. Changes must be pulled from, and pushed to, the same >>>> repository. (Pushes must still be done via ssh, not http.) If you >>>> have recorded "-gate" repository paths in any .hg/hgrc files then >>>> you'll have to remove the "-gate" suffix from those paths. >>>> >>>> - The unsupported Forest extension is no longer present on the servers, >>>> so commands such as fpush, fpull, fincoming, etc., will no longer >>>> work. As an alternative you can use the recently-published Trees >>>> extension [1] or the existing get_source.sh script [2]. >>>> >>>> - The minimum recommended version of Mercurial on the client side is >>>> 2.6.3, although earlier versions are likely to continue to work. >>>> >>>> - The ancient version of hg in /java/devtools will not be updated at >>>> this time, but will likely still work. That version is too old to >>>> support the Trees extension [1]. >>>> >>>> Additional note: >>>> >>>> - The SSH host keys on the cr, hg, and closedjdk servers have changed, >>>> so you'll need to update your known_hosts file. On Unix-like systems >>>> you can do that with these commands: >>>> >>>> $ ssh-keygen -R cr.openjdk.java.net >>>> $ ssh-keygen -R hg.openjdk.java.net >>>> $ ssh-keygen -R closedjdk >>>> $ ssh-keygen -R closedjdk.us.oracle.com >>>> >>>> Please report any problems to ops at openjdk dot java dot net. >>>> >>>> - Mark >>>> >>>> >>>> [1] http://openjdk.java.net/projects/code-tools/trees >>>> [2] http://openjdk.java.net/guide/repositories.html#clone >>> >> > > From daniel.fuchs at oracle.com Mon Mar 17 09:38:09 2014 From: daniel.fuchs at oracle.com (Daniel Fuchs) Date: Mon, 17 Mar 2014 10:38:09 +0100 Subject: CODETOOLS-7900327: Change webrev to use 'trees' instead of 'forest' Message-ID: <5326C281.4020903@oracle.com> Hi, I would like to propose a patch for fixing CODETOOLS-7900327: Change webrev to use 'trees' instead of 'forest' https://bugs.openjdk.java.net/browse/CODETOOLS-7900327 webrev: http://cr.openjdk.java.net/~dfuchs/webrev_CODETOOLS-7900327/webrev.00/ (I will need a sponsor) This patch changes the behavior of the '-f' option, so that it relies on 'trees' [1] instead of 'forest'. For this to work your forest needs to have been configured for trees. This happens automatically if you used 'hg tclone' to clone it, but otherwise you can call 'hg tconfig --set --walk --depth' to configure it (you can use 'hg tlist' to see the list of trees that have been configured). The main changes in the patch (trickiest) comes from the fact that whereas hg f used to print '[]' hg t now prints '[]:' I tested the patch with the following options: [all files edited, no commit]: webrev -fN [all files committed, no edits]: webrev -f best regards, -- daniel [1] http://openjdk.java.net/projects/code-tools/trees/ > Hi Daniel; > > Since it looks like several people have tried your patch and seem to find it works for them then perhaps you should just create a CODETOOLS issue and propose it to this list (webrev-dev) for sponsorship. > > I haven't had a chance to try it yet but likely wouldn't have anything to add by my review (and I am not a committer/reviewer on CODETOOLS) so I would encourage you to go ahead with proposing it yourself. > > Cheers, > > Mike From roger.riggs at oracle.com Mon Mar 17 13:05:25 2014 From: roger.riggs at oracle.com (roger riggs) Date: Mon, 17 Mar 2014 09:05:25 -0400 Subject: CODETOOLS-7900327: Change webrev to use 'trees' instead of 'forest' In-Reply-To: <5326C281.4020903@oracle.com> References: <5326C281.4020903@oracle.com> Message-ID: <5326F315.8050602@oracle.com> Hi Daniel, What error messages are generated if the tree extension is not configured? Should there be something in the usage to advise the correct preconditions? Roger On 3/17/2014 5:38 AM, Daniel Fuchs wrote: > Hi, > > I would like to propose a patch for fixing > > CODETOOLS-7900327: Change webrev to use 'trees' instead of 'forest' > https://bugs.openjdk.java.net/browse/CODETOOLS-7900327 > > webrev: > http://cr.openjdk.java.net/~dfuchs/webrev_CODETOOLS-7900327/webrev.00/ > > (I will need a sponsor) > > This patch changes the behavior of the '-f' option, so that it relies > on 'trees' [1] instead of 'forest'. > For this to work your forest needs to have been configured for trees. > This happens automatically if you used 'hg tclone' to clone it, > but otherwise you can call 'hg tconfig --set --walk --depth' to > configure it (you can use 'hg tlist' to see the list of trees > that have been configured). > > The main changes in the patch (trickiest) comes from the > fact that whereas hg f used to print '[]' > hg t now prints '[]:' > > > I tested the patch with the following options: > > [all files edited, no commit]: > > webrev -fN > > [all files committed, no edits]: > > webrev -f > > best regards, > > -- daniel > [1] http://openjdk.java.net/projects/code-tools/trees/ > >> Hi Daniel; >> >> Since it looks like several people have tried your patch and seem to >> find it works for them then perhaps you should just create a >> CODETOOLS issue and propose it to this list (webrev-dev) for >> sponsorship. >> >> I haven't had a chance to try it yet but likely wouldn't have >> anything to add by my review (and I am not a committer/reviewer on >> CODETOOLS) so I would encourage you to go ahead with proposing it >> yourself. >> >> Cheers, >> >> Mike > > > From daniel.fuchs at oracle.com Mon Mar 17 14:40:14 2014 From: daniel.fuchs at oracle.com (Daniel Fuchs) Date: Mon, 17 Mar 2014 15:40:14 +0100 Subject: CODETOOLS-7900327: Change webrev to use 'trees' instead of 'forest' In-Reply-To: <5326F315.8050602@oracle.com> References: <5326C281.4020903@oracle.com> <5326F315.8050602@oracle.com> Message-ID: <5327094E.8090708@oracle.com> Hi Roger, Here is a new webrev. It attempts to detect whether the trees extension is present by calling 'hg help trees' - and assuming that a non-zero error status means that trees are not configured. Also - I modified the places where I added calls to hgroot=`hg root` by hgroot=`hg root | $FILTER` as it seems it would be needed for cygwin. http://cr.openjdk.java.net/~dfuchs/webrev_CODETOOLS-7900327/webrev.01/ Note: the new additions (webrev.01) have been tested only on Mac. It would be good if someone could try this patch on windows/cygwin. I'm not sure anybody did that yet. best regards, -- daniel On 3/17/14 2:05 PM, roger riggs wrote: > Hi Daniel, > > What error messages are generated if the tree extension is not configured? > Should there be something in the usage to advise the correct preconditions? > > Roger > > On 3/17/2014 5:38 AM, Daniel Fuchs wrote: >> Hi, >> >> I would like to propose a patch for fixing >> >> CODETOOLS-7900327: Change webrev to use 'trees' instead of 'forest' >> https://bugs.openjdk.java.net/browse/CODETOOLS-7900327 >> >> webrev: >> http://cr.openjdk.java.net/~dfuchs/webrev_CODETOOLS-7900327/webrev.00/ >> >> (I will need a sponsor) >> >> This patch changes the behavior of the '-f' option, so that it relies >> on 'trees' [1] instead of 'forest'. >> For this to work your forest needs to have been configured for trees. >> This happens automatically if you used 'hg tclone' to clone it, >> but otherwise you can call 'hg tconfig --set --walk --depth' to >> configure it (you can use 'hg tlist' to see the list of trees >> that have been configured). >> >> The main changes in the patch (trickiest) comes from the >> fact that whereas hg f used to print '[]' >> hg t now prints '[]:' >> >> >> I tested the patch with the following options: >> >> [all files edited, no commit]: >> >> webrev -fN >> >> [all files committed, no edits]: >> >> webrev -f >> >> best regards, >> >> -- daniel >> [1] http://openjdk.java.net/projects/code-tools/trees/ >> >>> Hi Daniel; >>> >>> Since it looks like several people have tried your patch and seem to >>> find it works for them then perhaps you should just create a >>> CODETOOLS issue and propose it to this list (webrev-dev) for >>> sponsorship. >>> >>> I haven't had a chance to try it yet but likely wouldn't have >>> anything to add by my review (and I am not a committer/reviewer on >>> CODETOOLS) so I would encourage you to go ahead with proposing it >>> yourself. >>> >>> Cheers, >>> >>> Mike >> >> >> > From jonathan.gibbons at oracle.com Tue Mar 18 01:26:44 2014 From: jonathan.gibbons at oracle.com (Jonathan Gibbons) Date: Mon, 17 Mar 2014 18:26:44 -0700 Subject: CODETOOLS-7900327: Change webrev to use 'trees' instead of 'forest' In-Reply-To: <5326C281.4020903@oracle.com> References: <5326C281.4020903@oracle.com> Message-ID: <5327A0D4.7050300@oracle.com> Daniel, I can commit the change for you, but I'm not familiar with the webrev code so I'd like to see some general approval from folk that are. -- Jon On 03/17/2014 02:38 AM, Daniel Fuchs wrote: > Hi, > > I would like to propose a patch for fixing > > CODETOOLS-7900327: Change webrev to use 'trees' instead of 'forest' > https://bugs.openjdk.java.net/browse/CODETOOLS-7900327 > > webrev: > http://cr.openjdk.java.net/~dfuchs/webrev_CODETOOLS-7900327/webrev.00/ > > (I will need a sponsor) > > This patch changes the behavior of the '-f' option, so that it relies > on 'trees' [1] instead of 'forest'. > For this to work your forest needs to have been configured for trees. > This happens automatically if you used 'hg tclone' to clone it, > but otherwise you can call 'hg tconfig --set --walk --depth' to > configure it (you can use 'hg tlist' to see the list of trees > that have been configured). > > The main changes in the patch (trickiest) comes from the > fact that whereas hg f used to print '[]' > hg t now prints '[]:' > > > I tested the patch with the following options: > > [all files edited, no commit]: > > webrev -fN > > [all files committed, no edits]: > > webrev -f > > best regards, > > -- daniel > [1] http://openjdk.java.net/projects/code-tools/trees/ > >> Hi Daniel; >> >> Since it looks like several people have tried your patch and seem to >> find it works for them then perhaps you should just create a >> CODETOOLS issue and propose it to this list (webrev-dev) for >> sponsorship. >> >> I haven't had a chance to try it yet but likely wouldn't have >> anything to add by my review (and I am not a committer/reviewer on >> CODETOOLS) so I would encourage you to go ahead with proposing it >> yourself. >> >> Cheers, >> >> Mike > > > From jonathan.gibbons at oracle.com Tue Mar 18 01:37:00 2014 From: jonathan.gibbons at oracle.com (Jonathan Gibbons) Date: Mon, 17 Mar 2014 18:37:00 -0700 Subject: CODETOOLS-7900327: Change webrev to use 'trees' instead of 'forest' In-Reply-To: <5327A0D4.7050300@oracle.com> References: <5326C281.4020903@oracle.com> <5327A0D4.7050300@oracle.com> Message-ID: <5327A33C.7090704@oracle.com> Daniel, There are still references to the forest extension and to fstatus in the comments that should probably be removed or updated. For example, 1365 # 1366 # forest extension is still being changed. For instance the output 1367 # of fstatus used to no prepend the tree path to filenames, but 1368 # this has changed recently. AWK code below does try to handle both 1369 # cases 1370 # 1379 # 1380 # There is a bug in the output of fstatus -aC on recent versions: it 1381 # inserts a space between the name of the tree and the filename of the 1382 # old file. e.g.: 1383 # Also, it looks like 1396-1411 are working around the bug described in that comment. -- Jon On 03/17/2014 06:26 PM, Jonathan Gibbons wrote: > Daniel, > > I can commit the change for you, but I'm not familiar with the webrev > code > so I'd like to see some general approval from folk that are. > > -- Jon > > On 03/17/2014 02:38 AM, Daniel Fuchs wrote: >> Hi, >> >> I would like to propose a patch for fixing >> >> CODETOOLS-7900327: Change webrev to use 'trees' instead of 'forest' >> https://bugs.openjdk.java.net/browse/CODETOOLS-7900327 >> >> webrev: >> http://cr.openjdk.java.net/~dfuchs/webrev_CODETOOLS-7900327/webrev.00/ >> >> (I will need a sponsor) >> >> This patch changes the behavior of the '-f' option, so that it relies >> on 'trees' [1] instead of 'forest'. >> For this to work your forest needs to have been configured for trees. >> This happens automatically if you used 'hg tclone' to clone it, >> but otherwise you can call 'hg tconfig --set --walk --depth' to >> configure it (you can use 'hg tlist' to see the list of trees >> that have been configured). >> >> The main changes in the patch (trickiest) comes from the >> fact that whereas hg f used to print '[]' >> hg t now prints '[]:' >> >> >> I tested the patch with the following options: >> >> [all files edited, no commit]: >> >> webrev -fN >> >> [all files committed, no edits]: >> >> webrev -f >> >> best regards, >> >> -- daniel >> [1] http://openjdk.java.net/projects/code-tools/trees/ >> >>> Hi Daniel; >>> >>> Since it looks like several people have tried your patch and seem to >>> find it works for them then perhaps you should just create a >>> CODETOOLS issue and propose it to this list (webrev-dev) for >>> sponsorship. >>> >>> I haven't had a chance to try it yet but likely wouldn't have >>> anything to add by my review (and I am not a committer/reviewer on >>> CODETOOLS) so I would encourage you to go ahead with proposing it >>> yourself. >>> >>> Cheers, >>> >>> Mike >> >> >> > From daniel.fuchs at oracle.com Tue Mar 18 13:57:21 2014 From: daniel.fuchs at oracle.com (Daniel Fuchs) Date: Tue, 18 Mar 2014 14:57:21 +0100 Subject: CODETOOLS-7900327: Change webrev to use 'trees' instead of 'forest' In-Reply-To: <5327A33C.7090704@oracle.com> References: <5326C281.4020903@oracle.com> <5327A0D4.7050300@oracle.com> <5327A33C.7090704@oracle.com> Message-ID: <532850C1.9050007@oracle.com> Hi Jon, On 3/18/14 2:37 AM, Jonathan Gibbons wrote: > Daniel, > > There are still references to the forest extension and to fstatus in the > comments > that should probably be removed or updated. > > For example, > > 1365 # > 1366 # forest extension is still being changed. For instance the output > 1367 # of fstatus used to no prepend the tree path to filenames, but > 1368 # this has changed recently. AWK code below does try to handle > both > 1369 # cases > 1370 # > > > 1379 # > 1380 # There is a bug in the output of fstatus -aC on recent > versions: it > 1381 # inserts a space between the name of the tree and the filename > of the > 1382 # old file. e.g.: > 1383 # > > Also, it looks like 1396-1411 are working around the bug described in > that comment. Well - I'm not feeling bold enough to try and change that code. I have tested the new webrev with renamed files, both in the same dir and in different dirs and in subrepos, and it seems to still do the right thing. Therefore - I have simply updated the comments so that someone who would want to attempt and cleanup this code would understand that part of it is historical technological debt dating from the 'forest' time: http://cr.openjdk.java.net/~dfuchs/webrev_CODETOOLS-7900327/webrev.02/ I hope that can be enough for now; if it ain't broke... ;-) best regards, -- daniel > > -- Jon > > > On 03/17/2014 06:26 PM, Jonathan Gibbons wrote: >> Daniel, >> >> I can commit the change for you, but I'm not familiar with the webrev >> code >> so I'd like to see some general approval from folk that are. >> >> -- Jon >> >> On 03/17/2014 02:38 AM, Daniel Fuchs wrote: >>> Hi, >>> >>> I would like to propose a patch for fixing >>> >>> CODETOOLS-7900327: Change webrev to use 'trees' instead of 'forest' >>> https://bugs.openjdk.java.net/browse/CODETOOLS-7900327 >>> >>> webrev: >>> http://cr.openjdk.java.net/~dfuchs/webrev_CODETOOLS-7900327/webrev.00/ >>> >>> (I will need a sponsor) >>> >>> This patch changes the behavior of the '-f' option, so that it relies >>> on 'trees' [1] instead of 'forest'. >>> For this to work your forest needs to have been configured for trees. >>> This happens automatically if you used 'hg tclone' to clone it, >>> but otherwise you can call 'hg tconfig --set --walk --depth' to >>> configure it (you can use 'hg tlist' to see the list of trees >>> that have been configured). >>> >>> The main changes in the patch (trickiest) comes from the >>> fact that whereas hg f used to print '[]' >>> hg t now prints '[]:' >>> >>> >>> I tested the patch with the following options: >>> >>> [all files edited, no commit]: >>> >>> webrev -fN >>> >>> [all files committed, no edits]: >>> >>> webrev -f >>> >>> best regards, >>> >>> -- daniel >>> [1] http://openjdk.java.net/projects/code-tools/trees/ >>> >>>> Hi Daniel; >>>> >>>> Since it looks like several people have tried your patch and seem to >>>> find it works for them then perhaps you should just create a >>>> CODETOOLS issue and propose it to this list (webrev-dev) for >>>> sponsorship. >>>> >>>> I haven't had a chance to try it yet but likely wouldn't have >>>> anything to add by my review (and I am not a committer/reviewer on >>>> CODETOOLS) so I would encourage you to go ahead with proposing it >>>> yourself. >>>> >>>> Cheers, >>>> >>>> Mike >>> >>> >>> >> > From jonathan.gibbons at oracle.com Tue Mar 18 15:21:01 2014 From: jonathan.gibbons at oracle.com (Jonathan Gibbons) Date: Tue, 18 Mar 2014 08:21:01 -0700 Subject: CODETOOLS-7900327: Change webrev to use 'trees' instead of 'forest' In-Reply-To: <532850C1.9050007@oracle.com> References: <5326C281.4020903@oracle.com> <5327A0D4.7050300@oracle.com> <5327A33C.7090704@oracle.com> <532850C1.9050007@oracle.com> Message-ID: <5328645D.9010605@oracle.com> On 03/18/2014 06:57 AM, Daniel Fuchs wrote: > Hi Jon, > > On 3/18/14 2:37 AM, Jonathan Gibbons wrote: >> Daniel, >> >> There are still references to the forest extension and to fstatus in the >> comments >> that should probably be removed or updated. >> >> For example, >> >> 1365 # >> 1366 # forest extension is still being changed. For instance the >> output >> 1367 # of fstatus used to no prepend the tree path to filenames, but >> 1368 # this has changed recently. AWK code below does try to handle >> both >> 1369 # cases >> 1370 # >> >> >> 1379 # >> 1380 # There is a bug in the output of fstatus -aC on recent >> versions: it >> 1381 # inserts a space between the name of the tree and the filename >> of the >> 1382 # old file. e.g.: >> 1383 # >> >> Also, it looks like 1396-1411 are working around the bug described in >> that comment. > > Well - I'm not feeling bold enough to try and change that code. > I have tested the new webrev with renamed files, both in the same dir > and in different dirs and in subrepos, and it seems to still do > the right thing. > > Therefore - I have simply updated the comments so that someone who > would want to attempt and cleanup this code would understand that > part of it is historical technological debt dating from the 'forest' > time: > > http://cr.openjdk.java.net/~dfuchs/webrev_CODETOOLS-7900327/webrev.02/ > > I hope that can be enough for now; if it ain't broke... ;-) > > best regards, > > -- daniel > >> >> -- Jon >> >> >> On 03/17/2014 06:26 PM, Jonathan Gibbons wrote: >>> Daniel, >>> >>> I can commit the change for you, but I'm not familiar with the webrev >>> code >>> so I'd like to see some general approval from folk that are. >>> >>> -- Jon >>> >>> On 03/17/2014 02:38 AM, Daniel Fuchs wrote: >>>> Hi, >>>> >>>> I would like to propose a patch for fixing >>>> >>>> CODETOOLS-7900327: Change webrev to use 'trees' instead of 'forest' >>>> https://bugs.openjdk.java.net/browse/CODETOOLS-7900327 >>>> >>>> webrev: >>>> http://cr.openjdk.java.net/~dfuchs/webrev_CODETOOLS-7900327/webrev.00/ >>>> >>>> (I will need a sponsor) >>>> >>>> This patch changes the behavior of the '-f' option, so that it relies >>>> on 'trees' [1] instead of 'forest'. >>>> For this to work your forest needs to have been configured for trees. >>>> This happens automatically if you used 'hg tclone' to clone it, >>>> but otherwise you can call 'hg tconfig --set --walk --depth' to >>>> configure it (you can use 'hg tlist' to see the list of trees >>>> that have been configured). >>>> >>>> The main changes in the patch (trickiest) comes from the >>>> fact that whereas hg f used to print '[]' >>>> hg t now prints '[]:' >>>> >>>> >>>> I tested the patch with the following options: >>>> >>>> [all files edited, no commit]: >>>> >>>> webrev -fN >>>> >>>> [all files committed, no edits]: >>>> >>>> webrev -f >>>> >>>> best regards, >>>> >>>> -- daniel >>>> [1] http://openjdk.java.net/projects/code-tools/trees/ >>>> >>>>> Hi Daniel; >>>>> >>>>> Since it looks like several people have tried your patch and seem to >>>>> find it works for them then perhaps you should just create a >>>>> CODETOOLS issue and propose it to this list (webrev-dev) for >>>>> sponsorship. >>>>> >>>>> I haven't had a chance to try it yet but likely wouldn't have >>>>> anything to add by my review (and I am not a committer/reviewer on >>>>> CODETOOLS) so I would encourage you to go ahead with proposing it >>>>> yourself. >>>>> >>>>> Cheers, >>>>> >>>>> Mike >>>> >>>> >>>> >>> >> > Daniel, I'll take a look at your updated webrev, and file an RFE for the potential cleanup. -- Jon From iris.clark at oracle.com Tue Mar 18 16:19:32 2014 From: iris.clark at oracle.com (Iris Clark) Date: Tue, 18 Mar 2014 09:19:32 -0700 (PDT) Subject: CODETOOLS-7900327: Change webrev to use 'trees' instead of 'forest' In-Reply-To: <5328645D.9010605@oracle.com> References: <5326C281.4020903@oracle.com> <5327A0D4.7050300@oracle.com> <5327A33C.7090704@oracle.com> <532850C1.9050007@oracle.com> <5328645D.9010605@oracle.com> Message-ID: <83c16fe7-5c45-4e8d-83bb-819dea0d55af@default> Hi, Daniel. Please send me an example of the new functionality you'd like to highlight. I'll include it on this page: http://openjdk.java.net/guide/webrevHelp.html Thanks, iris -----Original Message----- From: Jonathan Gibbons Sent: Tuesday, March 18, 2014 8:21 AM To: Daniel Fuchs; webrev-dev at openjdk.java.net Subject: Re: CODETOOLS-7900327: Change webrev to use 'trees' instead of 'forest' On 03/18/2014 06:57 AM, Daniel Fuchs wrote: > Hi Jon, > > On 3/18/14 2:37 AM, Jonathan Gibbons wrote: >> Daniel, >> >> There are still references to the forest extension and to fstatus in >> the comments that should probably be removed or updated. >> >> For example, >> >> 1365 # >> 1366 # forest extension is still being changed. For instance the >> output >> 1367 # of fstatus used to no prepend the tree path to filenames, but >> 1368 # this has changed recently. AWK code below does try to handle >> both >> 1369 # cases >> 1370 # >> >> >> 1379 # >> 1380 # There is a bug in the output of fstatus -aC on recent >> versions: it >> 1381 # inserts a space between the name of the tree and the filename >> of the >> 1382 # old file. e.g.: >> 1383 # >> >> Also, it looks like 1396-1411 are working around the bug described in >> that comment. > > Well - I'm not feeling bold enough to try and change that code. > I have tested the new webrev with renamed files, both in the same dir > and in different dirs and in subrepos, and it seems to still do the > right thing. > > Therefore - I have simply updated the comments so that someone who > would want to attempt and cleanup this code would understand that part > of it is historical technological debt dating from the 'forest' > time: > > http://cr.openjdk.java.net/~dfuchs/webrev_CODETOOLS-7900327/webrev.02/ > > I hope that can be enough for now; if it ain't broke... ;-) > > best regards, > > -- daniel > >> >> -- Jon >> >> >> On 03/17/2014 06:26 PM, Jonathan Gibbons wrote: >>> Daniel, >>> >>> I can commit the change for you, but I'm not familiar with the >>> webrev code so I'd like to see some general approval from folk that >>> are. >>> >>> -- Jon >>> >>> On 03/17/2014 02:38 AM, Daniel Fuchs wrote: >>>> Hi, >>>> >>>> I would like to propose a patch for fixing >>>> >>>> CODETOOLS-7900327: Change webrev to use 'trees' instead of 'forest' >>>> https://bugs.openjdk.java.net/browse/CODETOOLS-7900327 >>>> >>>> webrev: >>>> http://cr.openjdk.java.net/~dfuchs/webrev_CODETOOLS-7900327/webrev. >>>> 00/ >>>> >>>> (I will need a sponsor) >>>> >>>> This patch changes the behavior of the '-f' option, so that it >>>> relies on 'trees' [1] instead of 'forest'. >>>> For this to work your forest needs to have been configured for trees. >>>> This happens automatically if you used 'hg tclone' to clone it, but >>>> otherwise you can call 'hg tconfig --set --walk --depth' to >>>> configure it (you can use 'hg tlist' to see the list of trees that >>>> have been configured). >>>> >>>> The main changes in the patch (trickiest) comes from the fact that >>>> whereas hg f used to print '[]' >>>> hg t now prints '[]:' >>>> >>>> >>>> I tested the patch with the following options: >>>> >>>> [all files edited, no commit]: >>>> >>>> webrev -fN >>>> >>>> [all files committed, no edits]: >>>> >>>> webrev -f >>>> >>>> best regards, >>>> >>>> -- daniel >>>> [1] http://openjdk.java.net/projects/code-tools/trees/ >>>> >>>>> Hi Daniel; >>>>> >>>>> Since it looks like several people have tried your patch and seem >>>>> to find it works for them then perhaps you should just create a >>>>> CODETOOLS issue and propose it to this list (webrev-dev) for >>>>> sponsorship. >>>>> >>>>> I haven't had a chance to try it yet but likely wouldn't have >>>>> anything to add by my review (and I am not a committer/reviewer on >>>>> CODETOOLS) so I would encourage you to go ahead with proposing it >>>>> yourself. >>>>> >>>>> Cheers, >>>>> >>>>> Mike >>>> >>>> >>>> >>> >> > Daniel, I'll take a look at your updated webrev, and file an RFE for the potential cleanup. -- Jon