PostgreSQL Weekly News - August 8, 2021

Posted on 2021-08-10 by PWN

PostgreSQL Weekly News - August 8, 2021

PGConf NYC is happening December 3-4, 2021. The CfP is open, as are opportunities to sponsor.

PostgreSQL Product News

Pgpool-II 4.2.4, 4.1.8, 4.0.15, 3.7.20 and 3.6.27, a connection pooler and statement replication system for PostgreSQL, released

pgmoneta 0.4.0, a backup and restore system for PostgreSQL, released

Buildfarm 13.1 software, a continuous integration system for the PostgreSQL project, released

dbForge Schema Compare 1.2 for PostgreSQL released

pg_timetable 4.0.0, a job scheduler for PostgreSQL, released.

Person of the week

PostgreSQL Jobs for August

PostgreSQL in the News

Planet PostgreSQL:

PostgreSQL Weekly News is brought to you this week by David Fetter

Submit news and announcements by Sunday at 3:00pm PST8PDT to

Applied Patches

Amit Kapila pushed:

Etsuro Fujita pushed:

  • Fix oversight in commit 1ec7fca8592178281cd5cdada0f27a340fb813fc. I failed to account for the possibility that when ExecAppendAsyncEventWait() notifies multiple async-capable nodes using postgres_fdw, a preceding node might invoke process_pending_request() to process a pending asynchronous request made by a succeeding node. In that case the succeeding node should produce a tuple to return to the parent Append node from tuples fetched by process_pending_request() when notified. Repair. Per buildfarm via Michael Paquier. Back-patch to v14, like the previous commit. Thanks to Tom Lane for testing. Discussion:

  • postgres_fdw: Fix issues with generated columns in foreign tables. postgres_fdw imported generated columns from the remote tables as plain columns, and caused failures like "ERROR: cannot insert a non-DEFAULT value into column "foo"" when inserting into the foreign tables, as it tried to insert values into the generated columns. To fix, we do the following under the assumption that generated columns in a postgres_fdw foreign table are defined so that they represent generated columns in the underlying remote table: * Send DEFAULT for the generated columns to the foreign server on insert or update, not generated column values computed on the local server.

  • Add to postgresImportForeignSchema() an option "import_generated" to include column generated expressions in the definitions of foreign tables imported from a foreign server. The option is true by default. The assumption seems reasonable, because that would make a query of the postgres_fdw foreign table return values for the generated columns that are consistent with the generated expression. While here, fix another issue in postgresImportForeignSchema(): it tried to include column generated expressions as column default expressions in the foreign table definitions when the import_default option was enabled. Per bug #16631 from Daniel Cherniy. Back-patch to v12 where generated columns were added. Discussion:

Andres Freund pushed:

Thomas Munro pushed:

Tom Lane pushed:

  • Doc: minor improvements for logical replication protocol documentation. Where appropriate, annotate message field data types with the backend code's name for the data type, eg XLogRecPtr or TimestampTz. Previously we just said "Int64" which didn't provide as much info. Also clarify references to object OIDs, and make use of the existing convention to denote the value of a field that must have a fixed value. Vignesh C, reviewed by Peter Smith and Euler Taveira. Discussion:

  • Add assorted new regexp_xxx SQL functions. This patch adds new functions regexp_count(), regexp_instr(), regexp_like(), and regexp_substr(), and extends regexp_replace() with some new optional arguments. All these functions follow the definitions used in Oracle, although there are small differences in the regexp language due to using our own regexp engine -- most notably, that the default newline-matching behavior is different. Similar functions appear in DB2 and elsewhere, too. Aside from easing portability, these functions are easier to use for certain tasks than our existing regexp_match[es] functions. Gilles Darold, heavily revised by me Discussion:

  • Don't elide casting to typmod -1. Casting a value that's already of a type with a specific typmod to an unspecified typmod doesn't do anything so far as run-time behavior is concerned. However, it really ought to change the exposed type of the expression to match. Up to now, coerce_type_typmod hasn't bothered with that, which creates gotchas in contexts such as recursive unions. If for example one side of the union is numeric(18,3), but it needs to be plain numeric to match the other side, there's no direct way to express that. This is easy enough to fix, by inserting a RelabelType to update the exposed type of the expression. However, it's a bit nervous-making to change this behavior, because it's stood for a really long time. (I strongly suspect that it's like this in part because the logic pre-dates the introduction of RelabelType in 7.0. The commit log message for 57b30e8e2 is interesting reading here.) As a compromise, we'll sneak the change into 14beta3, and consider back-patching to stable branches if no complaints emerge in the next three months. Discussion:

  • Really fix the ambiguity in REFRESH MATERIALIZED VIEW CONCURRENTLY. Rather than trying to pick table aliases that won't conflict with any possible user-defined matview column name, adjust the queries' syntax so that the aliases are only used in places where they can't be mistaken for column names. Mostly this consists of writing "alias.*" not just "alias", which adds clarity for humans as well as machines. We do have the issue that "SELECT alias.*" acts differently from "SELECT alias", but we can use the same hack ruleutils.c uses for whole-row variables in SELECT lists: write "alias.*::compositetype". We might as well revert to the original aliases after doing this; they're a bit easier to read. Like 75d66d10e, back-patch to all supported branches. Discussion:

  • Make regexp engine's backref-related compilation state more bulletproof. Up to now, we remembered the definition of a capturing parenthesis subexpression by storing a pointer to the associated subRE node. That was okay before, because that subRE didn't get modified anymore while parsing the rest of the regexp. However, in the wake of commit ea1268f63, that's no longer true: the outer invocation of parseqatom() feels free to scribble on that subRE. This seems to work anyway, because the states we jam into the child atom in the "prepare a general-purpose state skeleton" stanza aren't really semantically different from the original endpoints of the child atom. But that would be mighty easy to break, and it's definitely not how things worked before. Between this and the issue fixed in the prior commit, it seems best to get rid of this dependence on subRE nodes entirely. We don't need the whole child subRE for future backrefs, only its starting and ending NFA states; so let's just store pointers to those. Also, in the corner case where we make an extra subRE to handle immediately-nested capturing parentheses, it seems like it'd be smart to have the extra subRE have the same begin/end states as the original child subRE does (s/s2 not lp/rp). I think that linking it from lp to rp might actually be semantically wrong, though since Spencer's original code did it that way, I'm not totally certain. Using s/s2 is certainly not wrong, in any case. Per report from Mark Dilger. Back-patch to v14 where the problematic patches came in. Discussion:

  • Fix use-after-free issue in regexp engine. Commit cebc1d34e taught parseqatom() to optimize cases where a branch contains only one, "messy", atom by getting rid of excess subRE nodes. The way we really should do that is to keep the subRE built for the "messy" child atom; but to avoid changing parseqatom's nominal API, I made it delete that node after copying its fields to the outer subRE made by parsebranch(). It seems that that actually worked at the time; but it became dangerous after ea1268f63, because that later commit allowed the lower invocation of parse() to return a subRE that was also pointed to by some v->subs[] entry. This meant we could wind up with a dangling pointer in v->subs[], allowing a later backref to misbehave, but only if that subRE struct had been reused in between. So the damage seems confined to cases like '((...))...(...\2'. To fix, do what I should have done before and modify parseqatom's API to make it possible for it to remove the caller's subRE instead of the callee's. That's safer because we know that subRE isn't complete yet, so noplace else will have a pointer to it. Per report from Mark Dilger. Back-patch to v14 where the problematic patches came in. Discussion:

  • Rethink regexp engine's backref-related compilation state. I had committer's remorse almost immediately after pushing cb76fbd7e, upon finding that removing capturing subexpressions' subREs from the data structure broke my proposed patch for REG_NOSUB optimization. Revert that data structure change. Instead, address the concern about not changing capturing subREs' endpoints by not changing the endpoints. We don't need to, because the point of that bit was just to ensure that the atom has endpoints distinct from the outer state pair that we're stringing the branch between. We already made suitable states in the parenthesized-subexpression case, so the additional ones were just useless overhead. This seems more understandable than Spencer's original coding, and it ought to be a shade faster too by saving a few state creations and arc changes. (I actually see a couple percent improvement on Jacobson's web corpus, though that's barely above the noise floor so I wouldn't put much stock in that result.) Also, fix the logic added by ea1268f63 to ensure that the subRE recorded in v->subs[subno] is exactly the one with capno == subno. Spencer's original coding recorded the child subRE of the capture node, which is okay so far as having the right endpoint states is concerned, but as of cb76fbd7e the capturing subRE itself always has those endpoints too. I think the inconsistency is confusing for the REG_NOSUB optimization. As before, backpatch to v14. Discussion:

  • Doc: remove bogus <indexterm> items. Copy-and-pasteo in 665c5855e, evidently. The 9.6 docs toolchain whined about duplicate index entries, though our modern toolchain doesn't. In any case, these GUCs surely are not about the default settings of these values.

David Rowley pushed:

  • Track a Bitmapset of non-pruned partitions in RelOptInfo. For partitioned tables with large numbers of partitions where queries are able to prune all but a very small number of partitions, the time spent in the planner looping over RelOptInfo.part_rels checking for non-NULL RelOptInfos could become a large portion of the overall planning time. Here we add a Bitmapset that records the non-pruned partitions. This allows us to more efficiently skip the pruned partitions by looping over the Bitmapset. This will cause a very slight slow down in cases where no or not many partitions could be pruned, however, those cases are already slow to plan anyway and the overhead of looping over the Bitmapset would be unmeasurable when compared with the other tasks such as path creation for a large number of partitions. Reviewed-by: Amit Langote, Zhihong Yu Discussion:

  • Allow ordered partition scans in more cases. 959d00e9d added the ability to make use of an Append node instead of a MergeAppend when we wanted to perform a scan of a partitioned table and the required sort order was the same as the partitioned keys and the partitioned table was defined in such a way that earlier partitions were guaranteed to only contain lower-order values than later partitions. However, previously we didn't allow these ordered partition scans for LIST partitioned table when there were any partitions that allowed multiple Datums. This was a very cheap check to make and we could likely have done a little better by checking if there were interleaved partitions, but at the time we didn't have visibility about which partitions were pruned, so we still may have disallowed cases where all interleaved partitions were pruned. Since 475dbd0b7, we now have knowledge of pruned partitions, we can do a much better job inside partitions_are_ordered(). Here we pass which partitions survived partition pruning into partitions_are_ordered() and, for LIST partitioning, have it check to see if any live partitions exist that are also in the new "interleaved_parts" field defined in PartitionBoundInfo. For RANGE partitioning we can relax the code which caused the partitions to be unordered if a DEFAULT partition existed. Since we now know which partitions were pruned, partitions_are_ordered() now returns true when the DEFAULT partition was pruned. Reviewed-by: Amit Langote, Zhihong Yu Discussion:

  • Remove unused function declaration. It appears that check_track_commit_timestamp was declared but has never been defined in our code base. Likely this is just leftover cruft from a development version of the original patch to add commit timestamps. Let's just remove the useless declaration. The inclusion of guc.h also seems surplus to requirements. Author: Andrey Lepikhov Discussion:

Bruce Momjian pushed:

Peter Geoghegan pushed:

Dean Rasheed pushed:

  • Fix division-by-zero error in to_char() with 'EEEE' format. This fixes a long-standing bug when using to_char() to format a numeric value in scientific notation -- if the value's exponent is less than -NUMERIC_MAX_DISPLAY_SCALE-1 (-1001), it produced a division-by-zero error. The reason for this error was that get_str_from_var_sci() divides its input by 10^exp, which it produced using power_var_int(). However, the underflow test in power_var_int() causes it to return zero if the result scale is too small. That's not a problem for power_var_int()'s only other caller, power_var(), since that limits the rscale to 1000, but in get_str_from_var_sci() the exponent can be much smaller, requiring a much larger rscale. Fix by introducing a new function to compute 10^exp directly, with no rscale limit. This also allows 10^exp to be computed more efficiently, without any numeric multiplication, division or rounding. Discussion:

  • Adjust the integer overflow tests in the numeric code. Formerly, the numeric code tested whether an integer value of a larger type would fit in a smaller type by casting it to the smaller type and then testing if the reverse conversion produced the original value. That's perfectly fine, except that it caused a test failure on buildfarm animal castoroides, most likely due to a compiler bug. Instead, do these tests by comparing against PG_INT16/32_MIN/MAX. That matches existing code in other places, such as int84(), which is more widely tested, and so is less likely to go wrong. While at it, add regression tests covering the numeric-to-int8/4/2 conversions, and adjust the recently added tests to the style of 434ddfb79a (on the v11 branch) to make failures easier to diagnose. Per buildfarm via Tom Lane, reviewed by Tom Lane. Discussion:

Fujii Masao pushed:

Peter Eisentraut pushed: