Fail on condense when image name is drgeo.image
Hi, When I use the stable P7, drgeo build fails. There is an obscure error in File>>nextVersion, with negative index which are not catch, although it looks like it should be. This same build was working on alpha P7 of June. The stack of the debugger is attached as well the PharoDebug.log. Does it sound familiar to any of you? Hilaire -- Dr. Geo http://drgeo.eu
Hi Hilaire, On Mon, 4 Mar 2019 at 22:18, Hilaire <hilaire@drgeo.eu> wrote:
Hi,
When I use the stable P7, drgeo build fails.
There is an obscure error in File>>nextVersion, with negative index which are not catch, although it looks like it should be.
This same build was working on alpha P7 of June.
The stack of the debugger is attached as well the PharoDebug.log.
Does it sound familiar to any of you?
There were some changes made to FileReference>>nextVersion late last year (it was broken with only .1 and .1.1 version numbers being created). It looks like drgeo.changes.new has uncovered another issue as #nextVersion returns drgeo.1.new, when it should probably return drgeo.changes.1.new. But I can't reproduce the problem you're seeing. Can you create a small script that reproduces the issue? Cheers, Alistair
The problem seems to happen when there is the following conditions, in the folder containing the image: - there is a drgeo file, in my situation it is a link to the DrGeo repository - the image name is drgeo.image, ... - PharoCondense is executed, then FileRefrence>>nextVersion seems confused by the drgeo file. I could recheck, but it will be evening late. ----- http://drgeo.eu -- Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html
Hi Hilaire, On Tue, 5 Mar 2019 at 15:09, HilaireFernandes <hilaire@drgeo.eu> wrote:
The problem seems to happen when there is the following conditions, in the folder containing the image:
- there is a drgeo file, in my situation it is a link to the DrGeo repository - the image name is drgeo.image, ... - PharoCondense is executed, then FileRefrence>>nextVersion seems confused by the drgeo file.
I could recheck, but it will be evening late.
I had a bit of trouble reading the stack traces, but it looked like the issue is with FileReference>>nextVersion. If that is the case: #nextVersion is self contained, so the easiest thing will be to trigger the problem, go in to the debugger and copy the pathString of the FileReference that causes the issue. I've attached a change-set that fixes #nextVersion so it returns the correct values for file names with dots in them, e.g. "file.name.ext". Could you check if it behaves as expected? Cheers, Alistair
Alistair, I checked, the error shows under this specific situation I described in my previous email. Yes, your code fixes this condition, the condense does not hit this error anymore. Thanks Hilaire Le 05/03/2019 à 17:59, Alistair Grant a écrit :
I've attached a change-set that fixes #nextVersion so it returns the correct values for file names with dots in them, e.g. "file.name.ext". Could you check if it behaves as expected?
-- Dr. Geo http://drgeo.eu
Hi Hilaire, On Tue, 5 Mar 2019 at 21:16, Hilaire <hilaire@drgeo.eu> wrote:
Alistair,
I checked, the error shows under this specific situation I described in my previous email.
I figured out the environment that caused the primtiveFailure and can see how my modified code avoids the issue.
Yes, your code fixes this condition, the condense does not hit this error anymore.
Great. I've refactored this quite a bit to make it more efficient and related methods more consistent. I'll be opening an issue and submitting a PR for it. Cheers, Alistair
On Tue, 5 Mar 2019 at 22:11, Alistair Grant <akgrant0710@gmail.com> wrote:
Hi Hilaire,
On Tue, 5 Mar 2019 at 21:16, Hilaire <hilaire@drgeo.eu> wrote:
Alistair,
I checked, the error shows under this specific situation I described in my previous email.
I figured out the environment that caused the primtiveFailure and can see how my modified code avoids the issue.
Yes, your code fixes this condition, the condense does not hit this error anymore.
Great. I've refactored this quite a bit to make it more efficient and related methods more consistent. I'll be opening an issue and submitting a PR for it.
For future reference: Issue: https://github.com/pharo-project/pharo/issues/2753 PR: https://github.com/pharo-project/pharo/pull/2754 Cheers, Alistair
Thanks Alistair. Great! I can now build DrGeo with the P7.0.1 image. I will check with the P7.0.2, but should be fine. Now the OS X DrGeo build behave much more gently, but I need to test a bit further. Hilaire Le 06/03/2019 à 08:53, Alistair Grant a écrit :
For future reference:
Issue: https://github.com/pharo-project/pharo/issues/2753 PR: https://github.com/pharo-project/pharo/pull/2754
-- Dr. Geo http://drgeo.eu
participants (3)
-
Alistair Grant -
Hilaire -
HilaireFernandes