[Bug] appendonlyam_handler.c:1790:25: error: ‘segfile_count’ may be used uninitialized

[avamingli]

Apache Cloudberry version

main branch commit: 644b4e307380954e986f048d0a7910a3063e524e

What happened

appendonlyam_handler.c: In function ‘appendonly_index_build_range_scan’:
appendonlyam_handler.c:1790:25: error: ‘segfile_count’ may be used uninitialized [-Werror=maybe-uninitialized]
 1790 |                         FreeAllSegFileInfo(seginfo, segfile_count);
      |                         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
aocsam_handler.c: In function ‘aoco_index_build_range_scan’:
aocsam_handler.c:1952:25: error: ‘segfile_count’ may be used uninitialized [-Werror=maybe-uninitialized]
 1952 |                         FreeAllAOCSSegFileInfo(seginfo, segfile_count);
      |                         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
cc1: all warnings being treated as errors

What you think should happen instead

No response

How to reproduce

config, make

Operating System

Ubuntu

Anything else

No response

Are you willing to submit PR?

• [ ] Yes, I am willing to submit a PR!

Code of Conduct

• [x] I agree to follow this project's Code of Conduct.

[avamingli]

make failed, and something bad may be behind the failure, because segfile_count is not assigned any more. It's assigned before as:

	/*
	 * When Parallel index build,there is no additional operation to update the number of tuples
	 * that supports this logic. Uniform processing is used here. 
	 */ 
	if (progress)
	{
		seginfo = GetAllFileSegInfo(heapRelation, snapshot, &segfile_count, NULL);
		for (int seginfo_no = 0; seginfo_no < segfile_count; seginfo_no++)
		{
			total_blockcount += seginfo[seginfo_no]->varblockcount;
		}
		pgstat_progress_update_param(PROGRESS_SCAN_BLOCKS_TOTAL, total_blockcount);
	}

git blame show that it's changed by https://github.com/apache/cloudberry/pull/897

@reshke @yjhjstz @x4m @my-ship-it @leborchuk

[jiaqizho]

Also if (!FIPS_mode_set)(in contrib/pgcrypto/openssl.c) will got an WError..

[yjhjstz]

pls merge https://github.com/apache/cloudberry/pull/986 asap.

[reshke]

Maybe we should configure CI with more strict compiler flags... LGTM #986

[yjhjstz]

/*
	 * When Parallel index build,there is no additional operation to update the number of tuples
	 * that supports this logic. Uniform processing is used here. 
	 */ 
	if (progress)
	{
		seginfo = GetAllFileSegInfo(heapRelation, snapshot, &segfile_count, NULL);
		for (int seginfo_no = 0; seginfo_no < segfile_count; seginfo_no++)
		{
			total_blockcount += seginfo[seginfo_no]->varblockcount;
		}
		pgstat_progress_update_param(PROGRESS_SCAN_BLOCKS_TOTAL, total_blockcount);
	}

@reshke this need to check, I will merge quick fix https://github.com/apache/cloudberry/pull/986. issue keep open here. @avamingli

[avamingli]

and something bad may be behind the failure, because segfile_count is not assigned any more. It's assigned before as:

	/*
	 * When Parallel index build,there is no additional operation to update the number of tuples
	 * that supports this logic. Uniform processing is used here. 
	 */ 
	if (progress)
	{
		seginfo = GetAllFileSegInfo(heapRelation, snapshot, &segfile_count, NULL);
		for (int seginfo_no = 0; seginfo_no < segfile_count; seginfo_no++)
		{
			total_blockcount += seginfo[seginfo_no]->varblockcount;
		}
		pgstat_progress_update_param(PROGRESS_SCAN_BLOCKS_TOTAL, total_blockcount);
	}

git blame show that it's changed by #897

@reshke @yjhjstz @x4m @my-ship-it @leborchuk

Could someone please review this?

[reshke]

Upstream does this different way: https://github.com/greenplum-db/gpdb-archive/blob/main/src/backend/access/aocs/aocsam_handler.c#L2335 looks like we need to cherry-pick block directory support and restore this logic https://github.com/greenplum-db/gpdb-archive/blob/main/src/backend/access/aocs/aocsam_handler.c#L2196 i will take another look soon