Unify behavior w.r.t. each table when generating separately and altogether

[ThomasRebeleKeebo]

I'm using the latest master (361661a). There seems to be a problem when generating the files individually vs generating them all-at-once. Their checksum is different.

Here an example to illustrate the problem. It seems that the checksum depends on the existing files when generating all files at once (the first run has checksum 609ac507b1962fb60c2465bebef7f047, but the following ones have 9f96d2e4699ffa80d59471b5d550e7b0).

$ test() { ./dbgen -f $@ -s 0.01 2>&1 | grep -v -e "Version" -e "Council"; echo $(wc lineorder.tbl) $(md5sum lineorder.tbl) }
$ rm *.tbl
$ test
60176 80937 5526185 lineorder.tbl 609ac507b1962fb60c2465bebef7f047 lineorder.tbl
$ test -T l
60175 80976 5526437 lineorder.tbl c32dc281f11876c1331630b1f97025c2 lineorder.tbl
$ test
60179 80940 5526437 lineorder.tbl 9f96d2e4699ffa80d59471b5d550e7b0 lineorder.tbl
$ test -T l
60175 80976 5526437 lineorder.tbl c32dc281f11876c1331630b1f97025c2 lineorder.tbl
$ test
60179 80940 5526437 lineorder.tbl 9f96d2e4699ffa80d59471b5d550e7b0 lineorder.tbl
$ test   
60179 80940 5526437 lineorder.tbl 9f96d2e4699ffa80d59471b5d550e7b0 lineorder.tbl
$ test -T l
60175 80976 5526437 lineorder.tbl c32dc281f11876c1331630b1f97025c2 lineorder.tbl

Checking what happens when I start from scratch:

$ rm *.tbl
$ test
60176 80937 5526185 lineorder.tbl 609ac507b1962fb60c2465bebef7f047 lineorder.tbl
$ test -T l
60175 80976 5526437 lineorder.tbl c32dc281f11876c1331630b1f97025c2 lineorder.tbl
$ test
60179 80940 5526437 lineorder.tbl 9f96d2e4699ffa80d59471b5d550e7b0 lineorder.tbl

And for completeness, when creating only the lineorder table at the beginning:

$ rm *.tbl
$ test -T l
60175 80976 5526437 lineorder.tbl c32dc281f11876c1331630b1f97025c2 lineorder.tbl
$ test
60179 80940 5526437 lineorder.tbl 9f96d2e4699ffa80d59471b5d550e7b0 lineorder.tbl

The content of the file is noticeably different:

9f96d2e4699ffa80d59471b5d550e7b0  lineorder.tbl
1|1|209|1552|9|19940925|1-URGENT|0|17|2471035|11507269|4|2372193|87213|2|19941105|TRUCK|
1|2|209|674|2|19940925|1-URGENT|0|36|5668812|11507269|9|5158618|94480|6|19941121|MAIL|
1|3|209|637|1|19940925|1-URGENT|0|8|1230104|11507269|10|1107093|92257|2|19941127|REG AIR|
1|4|209|22|10|19940925|1-URGENT|0|28|2581656|11507269|9|2349306|55321|6|19941222|AIR|

vs

c32dc281f11876c1331630b1f97025c2  lineorder.tbl
1|1|74|1552|9|19960102|5-LOW|0|17|2471035|17279949|4|2372193|87213|2|19960212|TRUCK|
1|2|74|674|2|19960102|5-LOW|0|36|5668812|17279949|9|5158618|94480|6|19960228|MAIL|
1|3|74|637|1|19960102|5-LOW|0|8|1230104|17279949|10|1107093|92257|2|19960305|REG AIR|
1|4|74|22|10|19960102|5-LOW|0|28|2581656|17279949|9|2349306|55321|6|19960330|AIR|

[eyalroz]

Interesting...

TBH, I have not tested this extensively as I was fixing various issues, since - at least in principle - I wasn't supposed to be touching the results at all. In fact, I did not even test the initial revision to ensure the checksums match up.

I would have liked to say I am getting right on this, to identify and fix the issue. However - as you may be aware, I haven't touched this code for several years; and as you may or may not be aware - there's a genocidal war going on here (Palestine/Israel) and I'm trying to devote time to organizing and participating in protest activities. So, it'll be difficult for me to squeeze this in. I will try to do it, though. Please poke me again in a while if you see there is no progress.

[eyalroz]

So, as far as I can tell, this has always been the case - as far back as I can checkout without things breaking.

Marking this as an enhancement request.