I am in favor of unique filenames because it will guarantee we avoid any problems like this in the future, but in most cases, as long as the names are unique *in the test suite*, that should be sufficient. So I don’t have any objections to Zachary’s approach of making the filenames unique enough.
Thanks, -Stella From: Aleksandr Urakov <aleksandr.ura...@jetbrains.com> Sent: Thursday, December 6, 2018 12:25 AM To: Zachary Turner <ztur...@google.com> Cc: reviews+d54942+public+2603ca548f36d...@reviews.llvm.org; Stella Stamenova <sti...@microsoft.com>; llvm-comm...@lists.llvm.org; Hafiz Abid Qadeer <abidh....@gmail.com>; Raphael Isemann <teempe...@gmail.com>; lldb-commits@lists.llvm.org; sani...@subpath.org; chi...@raincode.com; Brian Gianforcaro <b.gia...@gmail.com> Subject: Re: [PATCH] D54942: [PDB] Make PDB lit tests use the new builder Yes, this solution would solve the current problem. But theoretically the uniqueness may be defined by the directory, where the file is located, not by the name. It looks very unlikely, but who knows... I thought about what Stella have said, and it seems that I have found a good solution for this. We can use something like `tempfile` (https://docs.python.org/3.7/library/tempfile.html<https://nam06.safelinks.protection.outlook.com/?url=https%3A%2F%2Fdocs.python.org%2F3.7%2Flibrary%2Ftempfile.html&data=02%7C01%7Cstilis%40microsoft.com%7Cbdbb5017a9b24a0ad26708d65b5458e0%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C636796815166605217&sdata=JXoPXUaLdantqlee9S0E7Huv8WIDoLHltXyn67GnABY%3D&reserved=0>) for every internally generated file of the script. It will guarantee us the uniqueness of the file. But for the uniqueness of files named by script parameters the script user will remain responsible. What do you think about the such approach? On Wed, Dec 5, 2018 at 11:55 PM Zachary Turner <ztur...@google.com<mailto:ztur...@google.com>> wrote: I was thinking that we could just automatically compute the output file names as: os.path.join(out_dir, basename(output_file) + '.' + basename(input_file) + '.obj') Currently it's just os.path.join(out_dir, basename(input_file) + '.obj') which is why I think the problem occurs. On Wed, Dec 5, 2018 at 12:47 PM Aleksandr Urakov <aleksandr.ura...@jetbrains.com<mailto:aleksandr.ura...@jetbrains.com>> wrote: With such solution there would be even no need to change the current commit. But I'm not sure that it's trivial to do - the output file name may contain path with directories. May be we can replace slashes with underscores in the output file path and concatenate it with the object file name? Or even replace slashes in the source file path and concatenate it with the output file path - so object files will be located in the same place as the output file. Am Mi., 5. Dez. 2018, 23:30 hat Zachary Turner <ztur...@google.com<mailto:ztur...@google.com>> geschrieben: It is not possible to specify object file name in compile and link mode. But perhaps we can just change the default object file name to include something from the output file as well On Wed, Dec 5, 2018 at 12:26 PM Aleksandr Urakov via Phabricator <revi...@reviews.llvm.org<mailto:revi...@reviews.llvm.org>> wrote: aleksandr.urakov added a subscriber: zturner. aleksandr.urakov added a comment. The similar problem with `typedefs.test` is here: http://lab.llvm.org:8014/builders/lldb-x64-windows-ninja/builds/1940/steps/test/logs/stdio<https://nam06.safelinks.protection.outlook.com/?url=http%3A%2F%2Flab.llvm.org%3A8014%2Fbuilders%2Flldb-x64-windows-ninja%2Fbuilds%2F1940%2Fsteps%2Ftest%2Flogs%2Fstdio&data=02%7C01%7Cstilis%40microsoft.com%7Cbdbb5017a9b24a0ad26708d65b5458e0%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C636796815166615226&sdata=CGSQScg8GAQnLurJBOCf1JaoT4hFhluoD6fVkvdECxw%3D&reserved=0> I have an assumption about the cause of the problem. Are the tests running in parallel? In this case `typedefs.test` and `enums-layout.test` are writing to the same object file together, because they both are compiled from the same source. @zturner Is it possible to specify object file's name in `compile-and-link` mode? Then we can specify different names in different tests for both object files and executables. But I think that splitting the source or combining the tests would be a better idea. I will fix it only tomorrow, because I'm already at home today. Feel free to revert it if needed. Repository: rL LLVM CHANGES SINCE LAST ACTION https://reviews.llvm.org/D54942/new/<https://nam06.safelinks.protection.outlook.com/?url=https%3A%2F%2Freviews.llvm.org%2FD54942%2Fnew%2F&data=02%7C01%7Cstilis%40microsoft.com%7Cbdbb5017a9b24a0ad26708d65b5458e0%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C636796815166625231&sdata=nagEt0RLn04uZdpSBsoisT4%2BxhlMLMK5vXYv3UkKdrg%3D&reserved=0> https://reviews.llvm.org/D54942<https://nam06.safelinks.protection.outlook.com/?url=https%3A%2F%2Freviews.llvm.org%2FD54942&data=02%7C01%7Cstilis%40microsoft.com%7Cbdbb5017a9b24a0ad26708d65b5458e0%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C636796815166625231&sdata=5O%2FLR7P8QEPgE7bhN0z9Szk0UgV1sChq8QNhAvzcHaU%3D&reserved=0> -- Aleksandr Urakov Software Developer JetBrains http://www.jetbrains.com<https://nam06.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.jetbrains.com&data=02%7C01%7Cstilis%40microsoft.com%7Cbdbb5017a9b24a0ad26708d65b5458e0%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C636796815166635240&sdata=upmrCJ165lcxyFu8cMo250E88LUpK8MG4G2ZjMpsp2A%3D&reserved=0> The Drive to Develop
_______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
