Thank you for your reply Martin! AUTO_PROFILE_VERSION should also be changed. Then create_gcov can be updated to support both the old format and the new format.
Eugene -----Original Message----- From: Martin Liška <mli...@suse.cz> Sent: Thursday, June 17, 2021 2:38 AM To: Eugene Rozenfeld <eugene.rozenf...@microsoft.com>; gcc-patches@gcc.gnu.org Subject: Re: [EXTERNAL] [PATCH] gcov: Use system IO buffering On 6/17/21 3:59 AM, Eugene Rozenfeld wrote: > |The commit from this patch > (https://nam06.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgcc.gnu.org%2Fgit%2F%3Fp%3Dgcc.git%3Ba%3Dcommit%3Bh%3D23eb66d1d46a34cb28c4acbdf8a1deb80a7c5a05&data=04%7C01%7CEugene.Rozenfeld%40microsoft.com%7C508d63026ea84be211cc08d9317395bb%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C637595194782996821%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=UG%2B41tXMZ94%2Ff80qCnmq%2BtZsFkLXc9NrdWF8KXwPjnk%3D&reserved=0) > changed the semantics of gcov_read_string and gcov_write_string. Before this > change the string storage was as described in gcov-io.h: "Strings are padded > with 1 to 4 NUL bytes, to bring the length up to a multiple of 4. The number > of 4 bytes is stored, followed by the padded string." After this change the > number before the string indicates the number of bytes (not words) and there > is no padding. Was this file format change intentional? Hello. Thanks for heads up! Yes, the change was intentional and I'm going to update documentation entry in gcov-io.h. > It breaks AutoFDO because create_gcov produces strings in the format > specified in gcov-io.h. Thanks, Eugene Sure, that needs to be adjusted. Martin
