Implement function to prune string keeping HTML closing tags

[aliksend]

Fixes #1587

[paskal]

Looks good! Could you please move the prune function to a different file? It's not telegram-specific and now increases complexity of telegram file without necessity.

[aliksend]

Sure. Sorry for the long wait

Moved function to separated file and rebased my branch onto current master

[paskal]

This is the only consonant to move to telegram.go, otherwise LGTM.

[aliksend]

you're right, my bad. moved back to telegram.go

[aliksend]

I don't know why build fails on CI/CD. Locally I can do both go build -ldflags "-X main.revision=0.0.0 -s -w" from backend/_example/memory_store directory and docker buildx build --build-arg SKIP_BACKEND_TEST=true --build-arg SKIP_FRONTEND_TEST=true -f backend/_example/memory_store/Dockerfile . from root directory. Can you help me with it?

[umputun]

I don't know why build fails on CI/CD. Locally I can do both go build -ldflags "-X main.revision=0.0.0 -s -w" from backend/_example/memory_store directory and docker buildx build --build-arg SKIP_BACKEND_TEST=true --build-arg SKIP_FRONTEND_TEST=true -f backend/_example/memory_store/Dockerfile . from root directory. Can you help me with it?

fixed, passed

[umputun]

I have generated some unit tests, as the one it had didn't seem too convincing to me.

[umputun]

pls note - the test is generated to match the code behavoir, however some of them look strange to me. for example this one

		{
 			name:      "text exceeding limit",
 			html:      "<p>Hello world, this is a long text</p>",
 			maxLength: 15,
 			expected:  "<p>...</p>",
 		},

I have not checked the code closely but the common sense expectation would be <p>Hello...</p> in this case.

@aliksend can you take a look pls and adjust tests/code if needed?

[aliksend]

Current implementation will generate string "<p>Hello ...</p>" which length is 16, exceeding the limit. But it make sense to remove space in cases like this.
Also I'll fix another tests to make behavior better.